TDS Light Curtain Profiler

"A Light Curtain and a Doppler Radar provide a Very Detailed Vehicle Profile."

The light curtain based profiler system provides a lateral matrix depiction of the vehicle.

The TDS light curtain based automatic vehicle classification system is a pattern recognition system that relies on special pattern recognition algorithms to categorize vehicles into a number of distinct types. It uses the length of the vehicle, the number and placement of axles and a lateral matrix depiction of the vehicle to form a complex pattern. This pattern is fed to discrimination software which correlates with one of a predetermined set of vehicle types.

The vehicle detection process is initiated when careful examination of the light curtain output indicates the presence of a vehicle. During periods when no vehicle is being detected by the light curtain, the processor periodically polls the light curtain looking for any indication of a vehicle presence. When the curtain is penetrated, the processor uses velocity information from the Doppler radar to determine the position of the vehicle relative to the initial location point set by the curtain. Based upon this position, the light curtain is polled at regular predefined distance intervals (as opposed to time). A complete set of data points is transmitted from the light curtain to the signal/data processor at each sample distance. From this information a vehicle profile is generated. This process continues until the light curtain no longer contains any indications of a vehicle. At that point the vehicle profile is terminated.

The placement of the lower elements of the curtain allow for the determination of axle distances from the front edge of the vehicle. The length of the vehicle is determined from the number of individual data sets that are recorded. Likewise the axle positions are determined from the data set information in the bottom beams. In severe weather environments, a treadle is added to provide the axle information during periods of inclement weather. A dual tire detector can also be added to the system in the event that dual tires play a role in the classification scheme.

The height of the light curtain is dependent on the classification requirements of the user. Simple axle based classification schemes can use a short curtain (3, 4 or 5 feet depending upon vehicle types). In order to provide for more detailed classes (trucks versus buses, motorcyles, et cetera), TDS increases the curtain height to 10 feet, using a curtain with 3/4 inch spacing in the lower five feet and 1.5 inch spacing in the upper five feet. The 3/4 inch spacing provides good hitch detection which the 1.5 inch upper beam spacing is sufficient for class differentiation.

Actual Profiles - Single Lane Light Curtain Based AVC System

Development History

1994 -
Transport Data Systems began the development of an AVC system using a ten foot light curtain and a Doppler radar. Using live traffic on the HOV exit lane on Route 163 in San Diego, the system was able to demonstrate a classification accuracy of greater than 99.5%.

1998 -
Transport Data Systems developed and tested an AVC system using a ten foot light curtain, a Doppler radar and a four strip treadle. Using live traffic on the New Jersey Turnpike, the system demonstrated a classification accuracy of greater than 99.5%. This work was performed for MFS Transportation Systems as part of the Regional Consortium project. The application was written to run under the Lynx real time operating system on the MFS lane controller. In addition to classifying vehicles into the New Jersey Turnpike and Garden State Parkway vehicle classes, the AVC system provides vehicle separation and vehicle location to the lane controller for transaction bounding and violation enforcement camera triggering.

1999 -
Transport Data Systems developed and tested an AVC system using a 3 foot light curtain, a Doppler radar and a four strip treadle. This system need only provide classification for cars, cars with trailers and buses. This work was performed for MFS Transportation Systems as part of the Regional Consortium project. The application was written to run under the Lynx real time operating system on the MFS lane controller. The system is used for classifying vehicles on the northern segment of the Garden State Parkway. In addition to classification of The system provides vehicle separation and vehicle location to the lane controller for transaction bounding and violation enforcement camera triggering.

TECHNICAL DESCRIPTIONS

Lateral Profilers